1.4.0

(description from #52)
Added a mean to get the timestamps of a given SMPTE Timecode. Two timebases are considered:

• Video system time: to_systemtime()
• Real time: to_realtime()

NTSC framerates will return different values as the system time is not aligned to the wall-clock time. All others should return the exact same value.

DF 29.97:

> tc = Timecode("29.97", '00:59:59;29')
> tc.to_systemtime()
'01:00:00.000'
> tc.to_realtime()
'00:59:59.996'

NDF 29.97:

> tc = Timecode("29.97", '00:59:59:29', force_non_drop_frame=True)
> tc.to_systemtime()
'01:00:00.000'
> tc.to_realtime()
'01:00:03.600'

PAL 50:

> tc = Timecode("50", '00:59:59:49')
> tc.to_systemtime()
'01:00:00.000'
> tc.to_realtime()
'01:00:00.000'

I did not implement setters as I don't think there's a meaningful usage or need to them.
Hopefully this may help others understand better Timecodes running at different rates and the relations to the wall-clock time.

The second commit is to preserve the drop frame flag on elementary operations.

> tc = Timecode(29.97, '00:00:00:00',  force_non_drop_frame=True)
> assert tc.drop_frame is False
> tc = tc + 1
> tc
'00:00:00:01' #before: '00:00:00;01'
> tc.drop_frame
False #before: True

Thanks for @cubicibo for contribution.

As usual, version 1.4.0 is also released to PyPI and running pip install timecode will install it.

1.3.2

No functional changes in this version, it only contains documentation update to docstring format and clarification on the accepted framerate values.

1.3.1

• Fix: Fixed 23.98, 29.97 DF, 29.97 NDF, 59.94 and 59.94 NDF rollover to 00:00:00:00 after 24 hours.

1.3.0

• Fix: Fixed a huge bug in 29.97 NDF and 59.97 NDF calculations introduced in v1.2.3.

• Fix: Fixed Timecode.framerate when it is given as 23.98. The framerate attribute will not be forced to 24 and it will stay 23.98.

• Update: Timecode.tc_to_frames() method now accepts Timecode instances with possibly different frame rates then the instance itself.

• Fix: Fixed Timecode.div_frames() method.

• Update: Test coverage has been increased to 100% (yay!)

1.2.5

• Fix: Fixed an edge case when two Timecodes are subtracted the resultant Timecode will always have the correct amount of frames. But it is not possible to have a Timecode with negative or zero frames as this is changed in 1.2.3.

• Fix: Fixed Timecode.float property for dropframe timecodes.

1.2.4

• Update: It is now possible to supply a Fraction instances for the framerate argument.

1.2.3

• Update: Passing frames=0 will now raise a ValueError. This hopefully will clarify the usage of the TimeCode as a duration. If there is no duration, hence the frames=0, meaning that the number of frames of the duration that this TimeCode represents is 0, which is meaningless.
• Update: Also added some validation for the frames property (oh yes it is a property now).

0.4.2

• Update: Version bump for PyPI.

0.4.1

• Fix: Fixed a test that was testing overloaded operators.

0.4.0

• New: Frame delimiter is now set to ":" for Non Drop Frame, ";" for Drop Frame and "." for millisecond based time codes. If Timecode.__init__() start_timecode is passed a string with the wrong delimiter it will be converted automatically.

• Update: All tests involving Drop Frame and millisecond time codes are now set to use the new delimiter.

• New: Timecode.tc_to_string() method added to present the correctly formatted time code.

• New: Timecode.ms_frame boolean attribute added.

• New: Timecode.__init__() now supports strings, ints and floats for the framerate argument.